Postoperative complications are associated with an impaired intestinal barrier function due to ischemic shock.1,2 The intestinal epithelial barrier is maintained mainly by intracellular tight junctions.3-5 These form a barrier between the apical and basolateral plasma membrane domains, regulating paracellular transport of molecules and ions. Recent literature suggests 2 important regulators of the tight junctions' assembly.6-10 The first is E-cadherin, a calcium-dependent molecule that promotes translocation of tight junction proteins to the plasma membrane and subsequent cell-cell adhesion. The second system is dependent on Na/K-ATPase, which is an ATPdependent system regulating Na/K exchange, thereby generating a transmembrane sodium gradient. Recently, it has been shown that E-cadherins are not sufficient to either re-establish epithelial polarity or suppress invasiveness.8 A combined expression of both E-cadherin and Na/K-ATPaseβ1 induced junctional complexes is associated with recovery of epithelial polarity and suppressed invasiveness.9

It has been suggested that dysfunctional formation of tight junctions after stressful events, such as elective surgery, contribute to postsurgical complications and delayed recovery.11-16 Several important systems are contributing to tight junction disruption. One of these systems involves the regulation of gut-blood flow.17-18 Decreased blood flow of the gut can result in mucosal damage by ATP depletion, free radical formation, and infiltration by neutrophilic granulocytes.17,18 The postoperative impairment of the epithelial barrier enables translocation of bacteria and their endotoxins. Enhanced bacterial translocation has been associated with increased postoperative complications such as multiple organ dysfunction.11,15,19-21

In cases of impaired epithelial intestinal barrier, a second system protects the body from invading pathogens and their endotoxins, namely the mononuclear phagocytic system (MPS), which is located predominantly in the liver.22-25 The MPS is essential for the clearance of endotoxins that may translocate from the gut. Failure of this second protective window gives pathogens the opportunity to invade organs other than the liver, thereby causing systemic infection.24-26 Ljungqvist et al have applied a hemorrhage and an endotoxemia stress model in the rat, and have suggested that hepatic liver glycogen content contributes to survival rate. Therefore, we hypothesized that liver glycogen contents contribute to the hepatic defense mechanism against infection. Also, glucose supplementation has been proven to increase gut blood flow.27,28

In surgery, where clamping of an artery (ischemia) is necessary to avoid excessive blood loss, the gut-blood flow is hampered during the reperfusion phase (removal of the clamp).16,17,24 Furthermore, fasting has a detrimental effect on liver MPS function.28

We hypothesized that presurgical administration of carbohydrates (CHO) not only reduces direct intestinal damage, but also protects more distal organs from infection by simultaneously maintaining the liver MPS function. To test this hypothesis, intestinal barrier function, bacterial translocation, and biochemical parameters were measured in an intestinal ischemia-reperfusion (I/R) rat model. The effect of a carbohydrate drinking solution, as an additive to the normal diet, was evaluated for these parameters.

Materials and Methods

Animals

The study was performed with male Wistar rats (200-300 g). Animals were subjected to an inversed day-night rhythm (9 PM-9 AM), 2 weeks before the surgical procedure. Rody weight and daily intake of food and fluids were determined. The animal experimental committee approved the animal experiments according to the Declaration of Helsinki.

Experimental Setup

Two groups of rats had access ad libitum to autoclaved chow feeding until 16 h before operation (I/R fasted =14, sham fasted = 4). Water remained ad lib. The third group had additional chow and water ad libitum access to a CHO solution (Table 1) in a drinking bottle, starting 5 days before operation and continuing until the operation (I/R CHO supplemented group =11). Chow was removed 16 hours before the operation.

On the day of operation, the animals were anesthetized by isoflurane iso-flow anesthesia O2/N^sub 2^O (3/1) with 2.5% vaporizing. Once the rats were under complete narcosis, the iso-flow anesthesia was set to 1.8% vaporizing for stable anesthesia. During the experiment, the rats were maintained at 37°C. The operation was a laparotomy followed by clamping of the superior mesenteric artery for 60 min followed by a reperfusion period of 180 min. After this period, blood was collected by heart puncture using ethylenediaminetetraacetic acid (EDTA) anticoagulation, followed by tying off a large section of ileum to prevent contamination in assessment of bacterial translocation. Subsequently, animals were killed. The intestine (ileum), liver, and kidney were collected and used for determination of the intestinal barrier function (ex vivo organ function), determination of bacterial translocation, or were immediately frozen in liquid nitrogen.

Determination of Intestinal Barrier Function

After collecting the ileum, the epithelium was stripped from the external muscle layer and mounted in Ussing chambers. After 15 minutes, 10^sup -5^ M horseradish peroxidase (HRP) was added to the mucosal compartment, and the mucosa] to serosal flux of HRP was determined at 4 time points (30, 60, 90, 120 minutes). The serosal concentration of enzymatically active HRP was measured using a method based on Gallati and Pracht.29

Bacterial translocation

The collected liver, kidney, spleen, and mesenteric lymph nodes (MLN) samples were homogenized in an Ultratorax. The homogenates at various dilutions were plated in triplicate on blood-reinforced Clostridia agar plates. These plates were incubated under anaerobic conditions at 37°C for 2-3 days, after which the bacteria were counted and calculated as colony forming units per gram wet tissue.

Measurement of blood urea

The urea concentration in blood plasma was determined using a kit (Sigma Diagnostics, No. 535). Data were expressed in millimolar values.

Determination of oxidative stress

Measurement of total glutathione (GSH + GSSG) was measured as described previously.30,31

Determination of the glycogen concentration in liver

Approximately 50 - 100 mg of tissue was freeze-dried. This freeze-dried tissue was powdered by use of a mortar. A total of 125 µL of 1 M HCl was added to 5 - 10 mg of freeze-dried tissue, without vortexing. The acidic hydrolysis was then allowed to proceed at 100°C for 2 h with occasionally mixing, resulting in generation of free glucose. Subsequently, the tubes were cooled on ice and thereafter 125 µL of 1 M NaOH was added. After mixing, tubes were centrifuged at 3,600 × g for 10 min at 4°C. The resulting supernatant was transferred into 2-mL tubes followed by centrifugation at 13,000 rpm in an Eppendorf centrifuge for 10 min. The glucose concentration was measured in the supernatant using a kit from Boehringer Mannheim (catalog no. 716251). Data are expressed as percentage of control.

Statistics

ANOVA was used for statistical analysis of the data. A P < .05 (2-sided) was considered significant. The Kolmogorov-Smirnov and the Shapiro-Wilk tests were applied to test for normality. The data were also tested by Levene's Test of Equality of Error Variance. The bacterial translocation data were not normally distributed and were therefore log-transformed, resulting in normal distribution. Correlations between the different parameters were calculated using the Pearson correlation coefficient. Data were analyzed using SPSS.

Results

Body weights were comparable between I/R fasted (249 ± 28 g) and I/R CHO (258 ± 29). Body weights of I/R sham operated animals were lower (215 ± 11).

To study the effects of I/R on the barrier function of the ileum, the HRP flux from the mucosal to serosal side was measured across stripped epithelium mounted in Ussing chambers. Previously, we have reported the effects of fasting and feeding on the susceptibility of the intestine to I/R damage. In this study, it was found that the HRP flux of control normal healthy rats was stable for 90 min (2 ± 2 µmol/cm^sup 2^/h).32 In the current study, the control group was sham-operated and preoperatively fasted. Figure 1 shows that results obtained with these shamoperated rats resemble our previous findings. Intestinal I/R in the fasted animals caused a significantly higher HRP flux from the mucosal to the serosal side than I/R in sham fasted animals (Figure 1). Rats that had preoperatively ad libitum access to CHO drinking solution showed a significantly lower flux of HRP across the epithelium of the ileum when compared with overnight fasted I/R rats (Figure 1).

Intestinal antioxidant capacity (total glutathione content) was analyzed to assess the severity of the oxidative stress generated by the intestinal I/R. The total glutathione content was significantly (P < .001) lowered in both the I/R fasted and the I/R CHO group (Figure 2) when compared with the sham fasted group. The difference observed between the I/R CHO and the I/R fasted animals was not statistically significant.

The bacterial content of liver, kidney, spleen, and mesenteric lymph nodes was measured (Figure 3). The I/R fasted rats showed a higher bacterial content in the liver, kidney, and mesenteric lymph nodes (Figure 3A-C) than sham fasted rats. The bacterial content of these organs (Figure 3A-C) in the I/R CHO group was significantly lower than in the I/R fasted animals. Bacterial content changes in the spleen were similar, but statistical significance was not reached.

It has been suggested that liver energy reserves might play a role in the severity of I/R injury. The liver energy reserves expressed in percentage glycogen content were approximately double in the I/R CHO group than in the I/R fasted group (Figure 4, 100% vs 50%). This difference reached statistical significance (P < .02).

The I/R fasted group showed a significantly (P < .01) higher blood plasma concentration of urea used as a clinical kidney function parameter (Figure 5) than the sham fasted group. In contrast, the I/R CHO-supplemented group showed a significantly lower blood plasma urea concentration than the I/R fasted animals (P < .05).

To evaluate whether bacterial contents and HRP flux are related, correlation factors between these parameters were calculated. No significant correlation was found between bacterial content of any of the organs and HRP flux. Nor did glycogen levels correlate with bacterial content of the different organs or HRP flux. Racterial content was, however, correlated with urea concentrations (Table 2).

Discussion

We tested the hypothesis that preoperative administration of CHO can protect the intestine from becoming dysfunctional in the postoperative state. In an intestinal I/R rat model, postoperative ileal HRP flux was measured as an indicator of intestinal barrier integrity. Sham operated fasted rats showed a low HRP flux comparable to normal healthy animals,33 suggesting that sham operation as well as overnight fasting had no influence on the intestinal barrier. I/R combined with fasting, however, resulted in a markedly decreased intestinal barrier function (Figure 1).34-38 Addition of a carbohydrate-rich drink significantly reduced this effect (Figure 1). Furthermore, a significant reduction in bacterial content of the liver, kidney, and MLN was shown. This might be a result of decreased bacterial translocation or increased bacterial clearance. The explanation in terms of decreased translocation is supported by reports on decreased adherence of bacteria to the intestinal wall due to preoperative nutrition.39,40

Alternatively, diminished bacterial translocation could be explained by improved MPS function of the liver, due to the increased glycogen storage. This is supported by the findings of Ljungqvist et al using a hemorrhage or an endotoxemia model for stress, who reported better survival rates in rats that were supplemented with CHO during the prestress period.41,42 Compared with fasted rats, CHO fed rats had an enhanced hepatic glycogen reloading capacity after hemorrhage, even when the fasted rats were given a glucose infusion after the hemorrhage to reach the same glucose plasma levels as the fed animals.41,42 From these experiments, it was concluded that the preoperative metabolic state is an important factor for preventing an abnormal postoperative metabolism. Presented data confirm these findings. Liver glycogen was significantly enhanced in preoperatively fed animals compared with fasted animals (Figure 4; P < .02). Higher glycogen levels in the liver might give a better position to cope with endotoxins, resulting in reduced bacterial translocation. Another explanation for reduced bacterial contents comes from literature suggesting that glucose administration improves gut-blood flow.27,28 Because bacterial translocation has also been reported as dependent upon transcellular transport over the intestine, improved intestinal gut flow might have consequences for bacterial defense capacity of the epithelial cell lining.43,44 Intestinal oxidative stress, measured as glutathione content, was not influenced by the glucose supplementation (Figure 2).

We searched for a relationship between barrier function and bacterial translocation (Table I). The correlation coefficients indicated that bacterial content was independent of intestinal barrier function. This finding is in accordance with findings that the clinic occurrence of sepsis is not directly correlated with intestinal permeability.45 Surprisingly, however, plasma urea did correlate with bacterial content. One could therefore speculate that the kidney is to some extent involved in bacterial clearance and that this function is hampered during fasting. The control of this experiment (ad lib water) was based on common practice in clinical hospitals. At this moment, countries apply either a strict fasting regime or have changed to new fasting guidelines, recommending clear fluids (water, coffee, tea) up to 2 h before surgery.46 The difference in body weight of both I/R groups compared with the sham fasted group might be explained by presence of edema in both I/R groups. CHO had no effect on this outcome. It cannot be completely excluded that differences in urea might be due to differences in fluid intake, although all fluids were presented ad libitum.

Our studies on a rat intestinal I/R model have previously shown that preoperative nutrition diminishes organ dysfunction and also prevents damage to organs like the heart.31,47 We hypothesized that this decrease in organ dysfunction was due to a decreased intestinal barrier function. Indeed, in the present and previous exploratory studies32 we found that HRP flux of I/R fasted animals was higher than that of control healthy rats, which supports the theory that multiple organ dysfunction is related to increased intestinal permeability. However, because the present data also suggest a better intestinal barrier to bacteria, independent of the barrier to molecules of the size of HRP, both mechanisms could contribute to the effect of preoperative CHO feeding. An improved molecular intestinal barrier would probably result in lower lipopolysaccharide levels leading to lower levels of inflammatory cytokines such as IL-6, TNF^sub α^, IFN^sub χ^, IL-2, and IL-4, while an improved barrier function to bacteria would result in decreased bacterial translocation.

In conclusion, preoperative intake of CHO results in a better-maintained intestinal barrier function and a higher liver glycogen concentration. These improvements in both intestine and liver might lead to the prevention of bacterial translocation and consequent systemic infection.